Eight Lessons from the Feedly Beta Process

A few friends responded to our previous post by asking what were the top lessons we learned so far from the feedly beta process. We are still very much heads down and have not really taken the time to step back and organize our thoughts but here is the raw data.

Lesson #1: Create a simple web site with a 2-min intro video. Your video was not the greatest but it helped people get a simple and quick tour of what feedly is. We got lucky enough that bwana came along and created his own intro video, which ended up being a lot better than ours. So try to reach out to bwana or find your bwana.

Lesson #2: You need a detonator. We were lucky enough to have been in contact with Louis Gray for a few months. Louis is pretty unique: he is able to play with the product and write detailed reviews, can be trusted and has the ears of a lot of A-listers. Thanks to Louis’ review, the news about feedly beta 1 started to spread like wild fire on friendfeed and twitter, reaching highly connected people like Leo Laporte, Chris Pirillo, Mashable, Webware, The Inquisitr,… So try to reach out to Louis Gray or find your Louis Gray.

Lesson #3: TechCrunch. If you want to use Techcrunch, make sure that they are briefed at the same time you are briefing everyone else and that they have enough time to write a post and announce the news at the same time everyone else does. In our case, given that we had been in contact with Louis for a few month and that he had played with previous versions of the product (and that he can write posts at lighting speed), we un-voluntarily killed the TechCrunch opportunity – which in hindsight did not end up being a bad thing because we would not have been able to handle their load correctly.

Lesson #4: Get a lot of sleep. In the old “software on a CD” world, you tend to push like crazy in the last few cycles to make sure that everything is as good as possible by the time it reaches the Gold master image. In the service world, turning on the beta switch is just the beginning of the journey. So make sure you get a *lot* of sleep the week before you plan to turn the switch on (see next point).

Lesson #5: Listen. Twitter and Friendfeed are your best friends. Get ready to get instant feedback minutes after you push the beta out. Thanks to twitter search and friendfeed search, you can get real-time candid feedback about your software. This is pure gold if you are ready to put your ego aside and listen. In a matter of minutes, you can see all the rough edges in the product pop out. In our case, we ran, during the first five days, into two really big issues (feedly not loading if third party cookies were not allowed and the integration between feedly and google reader not being explicit enough). In the old “software on a CD” world, those errors could have been fatal/extremely costly because it would have taken too long for us to get the feedback and adjust the problem. But thanks to twitter and friendfeed, we were able to identify both issues after just a couple hours and work with some of the users we angered the most to design and integrate the right fixes. The first five days ended up being insanely intense, with very little sleep and a lot of fire fighting – and all this is happening in an environment you have absolutely no control: you can not control what people say to each other over twitter. You can just listen and try to fix the product. So just make sure that you get a lot of sleep so that you can try ao absorbe as many of those punches as possible.

Lesson #6: Take care of your users. Get Satisfaction is a nice tool for capturing the frustration of your users and try to work with them to polish the rough edges. So take a look at it.

Lesson #7: Measure everything. One of the benefit of software as a service is that the service can continuously evolve. Some changes will be good. Some changes will be bad. Use Google Analytics Custom Events to monitor anonymously and non-intrusively which features are being used and which features are not being used. Usage is the single best indicator of whether an idea is good or not.

Lesson #8: Launch a Monday or a Tuesday. If you are lucky, the news will start to spread and at some point be picked up by some of the bigger blogs (Mashable, TechCrunch, Read Write Web, Life Hacker, etc..). Once the news hits one of them, it will start to spread internationally as well with blog taking part of those articles, translating them and adding some of their magic. The feedly news started on Monday on Louis’ blog and within 4 days, thanks to a lot of spontaneous tweets and blog posts, we had users in 142 countries. The signal seems to definitely decrease over the week ends so you probably want to have that entire process happen in a single week, hence the suggestion to announce early in the week.

We definitely did a lot of mistakes, learned a lot and had lots of fun during the process. These are the eight things that come to mind now. I will complete this list if other things come to mind.

Would you tend to agree these lessons? Did you recently launch a beta process and have interesting stories to share?

Author: @feedly

Read more. Know more.